linux-kernel - RE: [PATCH] Enter safe mode only when pins need to be reconfigured

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <CO1PR11MB483551294EC9A0E1FF2C2BFF96709@CO1PR11MB4835.namprd11.prod.outlook.com>
Date:   Tue, 23 Aug 2022 12:45:41 +0000
From:   "Khandelwal, Rajat" <rajat.khandelwal@...el.com>
To:     "Lee, Shawn C" <shawn.c.lee@...el.com>,
        "heikki.krogerus@...ux.intel.com" <heikki.krogerus@...ux.intel.com>,
        Greg KH <gregkh@...uxfoundation.org>
CC:     "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH] Enter safe mode only when pins need to be reconfigured

+Forgot to add Greg

-----Original Message-----
From: Lee, Shawn C <shawn.c.lee@...el.com> 
Sent: Monday, August 22, 2022 11:10 AM
To: heikki.krogerus@...ux.intel.com
Cc: linux-usb@...r.kernel.org; linux-kernel@...r.kernel.org; Khandelwal, Rajat <rajat.khandelwal@...el.com>; Lee, Shawn C <shawn.c.lee@...el.com>
Subject: [PATCH] Enter safe mode only when pins need to be reconfigured

There is no point to enter safe mode during DP/TBT configuration
if the DP/TBT was already configured in mux. This is because safe
mode is only applicable when there is a need to reconfigure the
pins in order to avoid damage within/to port partner.

1. if HPD interrupt arrives and DP mode was already configured,
safe mode is entered again which is not desired.
2. in chrome systems, IOM/mux is already configured before OS
comes up. Thus, when driver is probed, it blindly enters safe
mode due to PD negotiations but only after gfx driver lowers
dp_phy_ownership, will the IOM complete safe mode and send
ack to PMC.
Since, that never happens, we see IPC timeout.

Hence, allow safe mode only when pin reconfiguration is not
required, which makes sense.

Signed-off-by: Rajat Khandelwal <rajat.khandelwal@...el.com>
Signed-off-by: Lee Shawn C <shawn.c.lee@...el.com>

---
 drivers/usb/typec/mux/intel_pmc_mux.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/typec/mux/intel_pmc_mux.c b/drivers/usb/typec/mux/intel_pmc_mux.c
index d238913e996a..4bf84466d1ff 100644
--- a/drivers/usb/typec/mux/intel_pmc_mux.c
+++ b/drivers/usb/typec/mux/intel_pmc_mux.c
@@ -432,6 +432,25 @@ static int pmc_usb_connect(struct pmc_usb_port *port, enum usb_role role)
 	return pmc_usb_command(port, msg, sizeof(msg));
 }
 
+static bool
+pmc_usb_mux_allow_to_enter_safe_mode(struct pmc_usb_port *port,
+				      struct typec_mux_state *state)
+{
+	if ((IOM_PORT_ACTIVITY_IS(port->iom_status, DP) ||
+	     IOM_PORT_ACTIVITY_IS(port->iom_status, DP_MFD)) &&
+	     state->alt &&
+	     state->alt->svid == USB_TYPEC_DP_SID)
+		return false;
+
+	if ((IOM_PORT_ACTIVITY_IS(port->iom_status, TBT) ||
+	     IOM_PORT_ACTIVITY_IS(port->iom_status, ALT_MODE_TBT_USB)) &&
+	     state->alt &&
+	     state->alt->svid == USB_TYPEC_TBT_SID)
+		return false;
+
+	return true;
+}
+
 static int
 pmc_usb_mux_set(struct typec_mux *mux, struct typec_mux_state *state)
 {
@@ -442,8 +461,13 @@ pmc_usb_mux_set(struct typec_mux *mux, struct typec_mux_state *state)
 	if (port->orientation == TYPEC_ORIENTATION_NONE || port->role == USB_ROLE_NONE)
 		return 0;
 
-	if (state->mode == TYPEC_STATE_SAFE)
+	if (state->mode == TYPEC_STATE_SAFE) {
+		if (!pmc_usb_mux_allow_to_enter_safe_mode(port, state))
+			return 0;
+
 		return pmc_usb_mux_safe_state(port);
+	}
+
 	if (state->mode == TYPEC_STATE_USB)
 		return pmc_usb_connect(port, port->role);
 
-- 
2.31.1